Two different fractal descriptions have been previously found adequate to describe the heterogeneity of regional flows within the heart and lung over a 200-fold range of voxel sizes; one was based on statistical self similarity, the other on a self-similar branching algorithm. The two methods gave predictions of the heterogeneity at the 100 microgram level in the heart that differed by a factor of two, but were both compatible with the available data, which is not sufficiently accurate. The goals of this project are to develop methods for analyzing experiments providing data covering the three-order of magnitude gap from the previous data to the level of prediction, and to resolve the disparity in two ways: firstly, to develop new methods of describing the vascular network that cover the whole range of observable data (about five to six orders of magnitude) and secondly, to develop as working hypotheses those fractal approaches that do fit the data over the whole range.